Tae-young Jung
AI Model Optimization And Tool Development Engineer
Seoul, South Korea
Summary
👤
Senior🎓
Top SchoolTae-young Jung is an AI model optimization and tool development engineer with nine years of experience transitioning from video codec expertise in H.264/H.265 to building production NPU SDKs, compilers, quantizers, and Linux drivers. Based in Seoul, he has led NPU SDK development and evaluated third-party NPU IPs while researching state-of-the-art quantization methods to bridge research and deployable inference stacks. He contributes to Arm NN by improving ONNX parsing and expanding operator support, demonstrating a knack for making ML runtimes more robust and crash-resistant. Tae-young’s background in image communication and signal processing (MS) and PhD work in video compression informs a systems-level approach to performance and precision trade-offs in edge AI. Notably, he combines low-level codec and driver experience with higher-level compiler and tooling skills, making him effective at optimizing models across the full stack.
9 years of coding experience
13 years of employment as a software developer
Master of Science (MS) Image Communication & Signal Processing, Master of Science (MS) Image Communication & Signal Processing at Hanyang UniversityBachelor of Science (BS) Computer Engineering, Bachelor of Science (BS) Computer Engineering at Hongik University
English, Korean
Stackoverflow
1reputation
977reached
0answers
1question
Github Skills (8)
neural-network10
machine-learning10
c-language10
cprogramming-language10
mysqli6
php6
tensorflow5
pytorch5
Programming languages (6)
Github contributions (5)
ARM-software/armnn
Oct 2019 - Nov 2019
Arm NN ML Software. The code here is a read-only mirror of https://review.mlplatform.org/admin/repos/ml/armnn
Role in this project:
ML Engineer
ML EngineerContributions:8 commits, 5 PRs, 9 comments in 21 days
Contributions summary:Tae-young primarily contributed to the Arm NN ML Software repository by fixing bugs and improving the ONNX parser. Their work included addressing serialization issues related to the `qsymm16` data type, matching shapes between initializers and tensors, and resolving a crash issue related to how the parser handles tensor information. Additionally, the user extended the ONNX parser to support TanH, Sigmoid, and LeakyRelu layers, demonstrating a focus on expanding the library's capabilities. Furthermore, they addressed potential crashes involving zero-dimension tensors and zero-filled tensors.
adminarmv8deep-learningneural-networksmachine-learning
Tee0125/pytorch-detector-models
Feb 2020 - Oct 2021
Contributions:63 commits, 58 pushes, 1 comment in 1 year 7 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.